1Chapter 51: Radiology of the Nasal Cavity and Paranasal Sinuses

S. James Zinreich, Kenneth D. Dolan

The radiologic examination is designed to give information that is complementary orsupplementary to the clinical findings. By themselves radiographic changes are non-specificand require correlation with historical and physical examination findings to ensure the greatestdiagnostic usefulness.

Three major forms of radiologic evaluation are available. Plain films serve as a surveyevaluation of the transparency, size, and wall integrity of the sinuses. Customarily, the plain-film examination consists of Waters', Caldwell's, and lateral views of the paranasal sinuses, aswell as a basal view. In special cases the posteroanterior and Towne projections are veryuseful supplementary views for obtaining specific information.

The overlapping structures and the limited resolution of the fine body outlinesprovided by plain films hamper this modality in the evaluation of the sphenoid andparticularly in the evaluation of the ethmoid sinuses. Thin-section, pleural directionaltomography, with its 1 mm planar displays, partially addresses this limitation by avoidingsuperimposition of structures and providing improved bone detail. However, the cost oftomography (two to three times the cost of plain films), the increased radiation dose, and theincreased time required to perform the examination limit the use of this modality.

Computed x-ray tomography (CT), with its inherent capacity to simultaneously displaybone, soft tissue, and air, enables the scanner to optimally evaluate the maxillofacial bonestructures; nasal, orbital, and intracranial soft tissues; and minute nasal and paranasal sinusairways. Further, the originally obtained direct CT image data may be reconstructed toprovide additional indirect orthogonal planes - multiplanar reconstruction - resulting inimproved perception. This process is further optimized with three-dimensional reconstructionof the CT data, which will continue to facilitate surgical planning as well as the surgery itself.

Magnetic resonance imaging (MRI) is a radiologic examination technique that does notrequire ionizing radiation but rather depends on radiofrequency signals generated by nuclearprotons in tissue or recovery times required after shift in position of the protons.Unfortunately, as of yet, this modality is limited in its display of bone. However, MRI isproving to be beneficial in the diagnosis of fungal sinusitis and in distinguishing squamouscell carcinoma from inflammatory disease.

Plain-Film Examination

The specific radiographic details of positioning and exposure are well described instandard manuals of radiographic methods (see Ballinger, 1982). This section will presentanalysis of each plain-film view to show the main sinus territory each view depicts. For

2illustrative purposes, Fig. 51-1 is the plain-film examination from a patient with extensive nasalpolyposis and complete opacity of the sinuses on the left side.

Waters' view

Waters' view is a posteroanterior projection along a central plane in the occipitomentalaxis. This position places the maxillary sinus above the petrous bones and allows one to analyzethe transparency of the maxillary sinuses. In Fig. 51-1, A, the left maxillary sinus is completelyopaque while the right side is clear. The nasal fossa contains a large polyp on the left side. Alarge choanal polyp projects into the nasopharynx on the left and is evident through the openmouth. The anterior, orbital, and lateral walls of the maxillary sinus are intact. Table 51-1 listssupplementary structures visualized.

Caldwell's view

Caldwell's view uses an occipitofrontal beam. Consequently, the frontal sinus and noseare in contact with the film holder so that magnification is minimal, and the view can be usedto obtain a frontal sinus template for surgical purposes. In Fig. 51-1, B, the left frontal sinus isopaque to such a degree that the sinus lamina dura is barely definable. Opacity of the leftethmoid sinus is present in comparison to the right. The left orbital floor cortex and lateralmaxillary wall are thicker, indicating slight blastic osteitis.

Lateral view

In the lateral view projection the central x-ray beam should pass through the anatomiccenter of the sinuses. Transparency of the frontal, ethmoid, and maxillary sinuses is difficult tojudge because of overlap. Opacity of the sphenoid sinus may be evident, as in Fig. 51-1, C. Thelarge choanal polyp is evident.

Basal view

In the basal view the patient is in a submental vertex position with the mandible andfrontal sinus area superimposed. Two representations of bone details are possible with thisprojection. If underexposed, it demonstrates the zygomatic arches. If normally exposed, theethmoid sinus walls and septum, sphenoid sinus walls, and lateral maxillary, lateral orbital, andgreater wing of sphenoid margins are evident.

Fig. 51-1, D, demonstrates opacity of the left ethmoid and sphenoid sinuses. Overlyingteeth and a cap on the left second mandibular molar obscure the left maxillary sinus opacity.

3Supplemental Plain Films

Towne's view

The Towne or Chamberlain-Towne projection is an anteroposterior projection in whichthe central x-ray beam is directed downward 40 degrees in relation to the cantho-meatal baseline.Used as a routine part of the skull examination for the posterior fossa and foramen magnum, theview also defines the upper surface of the maxillary sinus alongside the inferior orbital fissure,and the auricular and subauricular parts of the mandible.

Posteroanterior view

The posteroanterior view makes use of an occipitofrontal central x-ray beam aligned withthe canthomeatal baseline. There the petrous ridges overlap the orbital roof on both sides. Thevalue of this projection lies in defining the ethmoid and sphenoid sinus roofs.

Conventional Tomographic Examination

Tomography using conventional radiographic techniques makes use of the blurring effectproduced when the x-ray tube and film are shifted in opposite directions during an exposure. Inthe plane of the locus of rotation no blurring occurs so that this layer is sharply defined whileeverything overlying and underlying that plane is blurred and not defined on the film. Thethickness of the sharp layer is inversely related to the length of tube and film excursion.Therefore the tomographic layer produced by a linear tube-film shift is relatively thick, about 5mm minimum, when compared with complex (hypocycloidal) motion that produces "cuts" ofabout 1 mm thickness. Complex motion also produces sharper edges in the tomographic planeand the most complete blurring of overlapping structures.

Generally, tomography is used to better visualize change within the sinus cavity or in thesinus wall when change cannot be thoroughly evaluated by plain-film examination.

Frontal tomograms are made in the anteroposterior position. Because of blurring ofpetrous bone, this area does not interfere with the image of the sinus being evaluated. In coronaltomograms one may visualize the cephalic, medial, lateral, and caudal margins of the sinus aswell as any soft tissue-air or fluid-air interface lying perpendicular to the tomographic plane.Structures that lie parallel to the tomographic plane are not well visualized; therefore a secondset of tomographic views is required to visualize the anterior or posterior edges of the sinus wallor any intrasinus abnormality. In some cases tomography of the suspicious area in an axial planemay be helpful.

Fig. 51-2, A, is a representative coronal section from the tomographic examination of apatient with biopsy-proven Wegener's granulomatosis. Plain films revealed opacity of the rightmaxillary sinus, ethmoid sinus, and nasal fossa. Bone erosion was suspected and demonstratedon this view, in which the right medial maxillary wall, maxillary-ethmoidal septum, and medial

4ethmoidal wall are missing. A soft tissue mass nearly fills the nasal fossa. Since visualizing theright lamina papyracea is difficult in this projection, an axial tomographic examination was alsodone (Fig. 51-2, B). The tomogram shows destruction of ethmoidal secondary septae but not ofthe lamina papyracea.

Even though the larger maxillofacial bone structures are better defined by tomographythan plain films, the blurring (phantom artifact) that superimposes the ethmoid air channelsprecludes accurate definition. Moreover, the blurring artifact can be mistaken for mucoperiostealdisease, resulting in an overestimation of inflammatory disease.

Computed Tomography

CT differs from conventional radiographic study in several ways. First, the x-ray tube isaligned with a detector crystal or banks of detectors rather than an x-ray film. A highlycollimated (very thin) x-ray beam is projected through the part being examined so that only a thinlayer of the patient is exposed for each slice, in contrast to conventional tomography, in whichthe entire volume being studied is exposed even though only a small layer of the volume is infocus on each section. Therefore exposure of the patient to x-rays is much smaller in a CTexamination than for conventional tomography. For example, if a 5 rad exposure is used for eachconventional tomogram and 10 slices are obtained, the patient receives a total exposure of 50 rad.If 5 rad were used for a similar 10-slice CT examination of the same area, exposure through theentire volume would amount to only 5 rad (a slight additive effect may occur if adjacent CTslices overlap). Currently, CT scanning parameters are being adjusted to further decrease theradiation exposure. By reducing the milliamperage (Ma) and time of scanning, the radiationexposure can be reduced to 1 or 2 rads.

The coronal plane is the plane closest to the view of the endoscopist; it is also theimaging plane that best displays the ostiomeatal unit. Thus it is the preferred plane for directscanning. Each patient is positioned prone with the head hyperextended on the scanner bed (Fig.51-3). For optimal visualization of the ostiomeatal channels, the field of view should be focusedon the paranasal sinuses (on the Siemens Somatom DR 3) scanner we used a zoom of 4 or 5).Scanner computation algorithms are selected to favor the demonstration of soft tissue. Windowwidths are usually at 2000, and the window is centered to -200. These parameters are optimalfor display of the regional anatomy and for the evaluation of patients with chronic inflammatorydisease. For the demonstration of adjacent pathologic conditions (in the face) a narrower windowrange is necessary. Scanner "raw" data are transiently saved so that high-resolution bone-enhancing reconstructions can be applied when bone erosion was either visualized or suspected.

When patients are unable to assume the prone position, axial scans from the palatethrough the frontal sinus are obtained and indirect coronal reconstructions are then generatedfrom them. For special attention to the anterior ethmoid region, coronal indirect reconstructionsare performed to complement the initial scanning plane. In our experience, even in patients withextensive metallic dental fillings, the direct coronal plane proved superior to indirectreconstructions, and therefore the coronal plane remains the plane of choice.

5Anatomy

Ethmoidal labyrinth

As seen on the coronal view, air cells collectively form the ethmoidal labyrinth. Theyappear as a near vertically oriented, thinly septated bony honeycomb lined by mucosa. Thesevertically situated air cells are narrower anteriorly and wider posteriorly. The boundaries of thislabyrinthine structure are the lamina papyracea laterally, the orbital plate of the frontal bonesuperiorly, the perpendicular plate medially, and the middle turbinate inferiorly.

Ostiomeatal unit

Maxillary sinus ostium and infundibulum

The maxillary sinus ostium and the infundibulum serve as the predominant channellinking the maxillary sinus with the nasal cavity. They are best visualized in the coronal plane.The infundibulum is bounded laterally by the inferomedial orbit, superiorly by the hiatussemilunaris and ethmoidal bulla, medially by the uncinate process, and inferiorly by the maxillarysinus as the sinus funnels into it. Less frequently, an accessory orifice of the maxillary sinusesis encountered. The accessory orifice most frequently opens into the anterior fontanel of the nasalcavity and is best seen on modified axial views, assuming a midposition between "true" axial andcoronal planes. The infundibulum represents the superomedial extension of the ostium. Theposterior extent of the uncinate process and the relative position of the ostium determine whetherthe ostium may be visualized on endoscopy (Fig. 51-4).

Hiatus semilunaris

This complex space gains its name from its arched appearance in the sagittal plane. Thehiatus semilunaris is bounded superiorly by the ethmoidal bulla, laterally by the medial bonyorbit, inferiorly by the uncinate process, and medially by the middle meatus. The hiatussemilunaris is the final segment for drainage from the maxillary sinus, being preceded by themaxillary ostium and infundibulum. The hiatus semilunaris is best identified on parasagittalsections and runs obliquely in a posteroinferior direction between the uncinate process and theethmoidal bulla. A posterosuperior extension of the hiatus semilunaris (hiatus semilunarissuperioris) passes between the ethmoidal bulla and basal lamella and communicates with thesinus lateralis, affording drainage for this space.

Middle turbinate

The middle turbinate lies inferomedial to the anterior ethmoid air cells. Its most consistentbony attachments are vertical to the cribriform plate superiorly and to the lamina papyracealaterally via a bony strut termed the basal (ground) lamella. The basal lamella is orientedanteromedially to posterolaterally to become situated behind the ethmoidal bulla. Thecompartment between the posterior wall of the ethmoidal bulla and the basal lamella is the sinus

6lateralis. Quite often the body of the middle turbinate contains an air-filled cavity, the conchabullosa, which communicates variably with the superior medial meatus, the frontal recess, or thesinus lateralis.

Ethmoidal bulla

The ethmoidal bulla usually consists of an air cell of variable size and shape. It isbordered inferomedially by the infundibulum and hiatus semilunaris, laterally by the laminapapyracea, and superoposteriorly by the sinus lateralis. It communicates with the nasal cavity viaan ostium, the site of which appears to be variable. According to Zuckerkandel (1893) the ostiumis most often posterior, but according to Messerklinger (1967) a superoanterior opening is morefrequent. We have also noted it medially.

Frontal recess

The frontal recess affords mucociliary drainage of the frontal sinus. Drainage may occurdirectly into the middle meatus medial to the uncinate process, into the ethmoidal infundibulummore laterally, or more posteriorly above the ethmoidal bulla. This communication between thefrontal sinus and the nasal cavity is not strictly a duct but an internal aperture of hourglassconfiguration positioned between the sinus and the anterior middle meatus.

Nasolacrimal duct

The nasolacrimal duct is a straight-coursing tube that extends upward from the lacrimalfossa to a site adjacent to the attachment of the inferior turbinate. In the coronal view the ductis nearly superoinferiorly oriented, with its inferior portion lying about 3 to 5 degrees medial toits superior portion. In the sagittal view its posterior incline may be larger, varying from 5 to 30degrees.

Sphenoidal sinus, sphenoid ostium, and sphenoethmoid recess

This continuum is best evaluated on either axial or sagittal scans. The ostium is locatedat the anterosuperior portion of the sphenoid sinus. The sphenoid ostium and the posteriorethmoidal air cells drain into the sphenoethmoidal recess.

Magnetic Resonance Imaging

Fig. 51-5 is an example of a magnetic resonance image made in the coronal plane thatpasses through the ethmoid and maxillary sinuses. Since no radio frequency signal is generatedin the sinus areas, they appear black. The inferior and middle turbinates, which are larger on theleft, produce a fairly intense signal. A very strong signal arises from the retrobulbar fat, in thecenter of which is a round black area representing the optic nerves. Medial to this is a verticalblack line representing the medial rectus muscle. No signal arises from the maxilla, mandible,or teeth. A weak signal arises from the tongue muscles.

7The nasal cycle is optimally demonstrated by MRI. The edematous soft tissue that occurscyclically in the nasal cavity assumes a high signal intensity and should not be confused with apathologic process. We have found that the mucosa of the nasal cavity, as well as the mucosaof the ethmoid sinus, takes part in this cyclic event. This cycle can vary from 20 minutes to 6hours in individual patients. The mucosa in the maxillary, frontal, and sphenoid sinuses does notcycle. Thus the effects of the cyclic process can easily interfere with the estimation ofinflammatory disease in the nasal cavity and ethmoid sinus. On the other hand, the presence ofany increased signal within the maxillary, frontal, and sphenoid sinuses may be considered apathologic process (Fig. 51-6).

MRI has proven very beneficial in its ability to distinguish between fungal and bacterialinfection. Bacterial and viral infections assume a high signal intensity on T2-weighted images,whereas fungal concretions have virtually no signal, similar to the signal intensity of air.Similarly, MRI is very helpful in distinguishing between neoplastic and inflammatory pathologicconditions. Squamous cell carcinomas (90% of tumors in this area) have an intermediate rise insignal intensity on T2-weighted images compared to the very strong increase in signal intensitymanifested by bacterial and viral infections.

Basic Abnormalities

Size

Sinus size depends on developmental factors and may also be influenced by acquiredextrinsic mass lesions. Variations in sinus size may affect sinus transparency if compressivechange is great. Transparency of a sinus also depends on the content of a sinus space and theintegrity of the sinus wall.

Considerable variation in sinus size occurs from patient to patient. A tendency towardabnormality seems to exist in those sinuses that are smaller than normal. This tendency isespecially evident in respect to the maxillary sinus; in our experience the symptomatic sinus isoften the smaller one.

Size governs sinus transparency on transillumination and conventional radiographicstudies. In general, a small sinus is less transparent than a normal one. Fig. 51-7, A, is Waters'view of a patient with right maxillary sinus hypoplasia. The diminutive transverse andperpendicular lengths of the right maxillary sinus differ sharply from the dimensions of theaverage-sized left sinus. The right maxillary sinus extends just past the right infraorbital canal,which is the usual size of the maxillary sinus in a 5-year-old child. The normal left sinus has azygomatic recess and a small alveolar recess inferiorly.

The basal view of this patient is shown in Fig. 51-7, B. The dental structures obscure theright maxillary sinus, but the perimeters of the left side are easily visualized. In hypoplasia ofthe maxillary sinus the medial sinus surface is farther from the bony septum, resulting inenlargement of the nasal fossa transverse dimension. This fossal enlargement is evident in both

8views of Fig. 51-7. The normal left maxillary sinus is much more transparent than the small rightcell. The lamina dura of the small right sinus is well defined and sharp, even though thethickness of the right lateral maxillary wall is greater than the left. The patient was asymptomaticin respect to the sinuses and was being evaluated for nasal injury.

Extreme variation in frontal sinus size may be found, as shown in Fig. 51-8. The patientin Fig. 51-8, A, has absence (aplasia) of the frontal sinuses. The anterior ethmoid cells producethe small, rounded air cells in the lower frontal area. By way of contrast, the air cells in Fig. 51-8, B, are hyperplastic and serpiginous. One portion of the left sinus extends into the orbitalprocess of the frontal bone. Another frontal sinus variation is seen in Fig. 51-8, C, in which twonasal ridge cells overlie the lower frontal sinus. These cells lie within the sinus and occasionallymay be large enough to obstruct the nasofrontal ducts, but usually they are just a variation andshould not be confused with a pathologic process.

Ethmoid sinuses are rarely small; more commonly, they are enlarged and extendthroughout the orbital roof as in Fig. 51-9, in which essentially all of the orbital roof ispneumatized on both sides. Orbital roof pneumatization may be unilateral.

Perhaps the most variation in size is found in the sphenoid sinus. Small cells that extendonly into the presphenoidal area or just under the sella may be found. More commonlyhyperplasia is present with recesses of the sphenoid sinus into the anterior and posterior clinoidprocess, lesser wings, and greater sphenoid wings. Fig. 51-10 illustrates sphenoid sinushyperplasia: lesser sphenoid wing recesses are shown in Fig. 51-10, A; large sphenoidal recessesextending into the greater wing are demonstrated in Fig. 51-10, B; and anterior and posteriorclinoid process recesses are shown in Fig. 51-10, C. In Fig. 51-10, D, a large sphenoidal air cellrecess is shown extending far laterally into the orbital process of the greater wing and couldsuggest a pathologic process.

In general, no relation to the variations in size among the sinuses, seems to exist.Hypoplasia of all sinuses occurs rarely. Sinus size is related to surrounding structures. Forexample, after removal of an injured eye in a child, significant increase in ethmoid sinus sizeoccurs unless a replacement ocular prosthesis is used. Similar increase in size of the frontal sinusmay be found in the patient with unilateral cerebral atrophy. Conversely, compressive decreasein sinus size may occur as a result of expansion by a nearby mass. Fig. 51-11, A, shows a normalleft frontal sinus. The right frontal bone has a hyperlucent area that resembles a large frontalsinus, but no definite sinus lamina dura is present. The lateral view (Fig. 51-11, B) illustratesmarked thinning of the right frontal bone. Similar thinning was present on a basal view. CTstudies in Fig. 51-11, C and D, confirm the thin right frontal bone and show an underlyingarachnoid cyst as the cause of thinning.

Compressive change in sinus size may result from many causes. A giant carotid aneurysmmay compress a sphenoid sinus cell. Lesions along the floor of the anterior fossa may compressthe ethmoidal cells. Neurofibromas along the mandibular nerve may be large enough to compressthe lateral maxillary sinus wall, and slow-growing masses arising within the pterygomaxillary

9area may compress the posterior maxilla. This led Potter (1969) to use a slightly rotated lateralview to separate the pterygomaxillary spaces on the two sides, thus uncovering unilateralenlargement that is a sign of juvenile nasal angiofibroma.

Air-fluid level

The prone Waters' view in Fig. 51-12, A, shows opacity of the right maxillary sinus. Thesame patient was examined in the upright position (Fig. 51-12, B), which revealed an air-fluidlevel on the right. Distribution of the fluid along the anterior sinus surface in A produced theappearance of complete opacity.

The lateral view may also reveal a maxillary air-fluid level, but the air-soft tissue interfaceof the nasal turbinate may resemble an air-fluid level. A maxillary air-fluid level is evident inFig. 51-12, C. CT examination defines the maxillary air-fluid level particularly well, as seen inFig. 51-12, D. Air-fluid levels are not evident in conventional radiographic examinations of theethmoid sinuses but may occasionally be seen as intracellular change on CT studies.

Frontal sinus and sphenoid sinus air-fluid levels are present in Fig. 51-13. If the amountof fluid present in a sinus space is very small, the overlying structures or a sinus margin mayobscure it. The presence of fluid in the sinuses is usually secondary to an acute inflammatoryprocess, the result of hemorrhage from trauma, or of iatrogenic origin.

Domed sinus mass

The air-fluid level typically has a straight or slight meniscus-shaped interface. Domedlesions have a convex interface with the air contained in the sinus. Most commonly these lesionsrepresent mucus or serous mucus-retention cysts. Blood clots and subperiosteal bloodaccumulations may occasionally be rounded. Benign or malignant tumors may have a domeshape. Dentigerous lesions, particularly cysts, will have a dome shape if the lesion projects intoa maxillary sinus. Fascenelli (1969) has shown that domed lesions are very common and can befound in 10% of a population group.

Mucus-retention cysts may have quite a variable appearance in the maxillary sinus, asdemonstrated in Fig. 51-14. In Fig. 51-14, A, a large cyst causes almost complete opacificationof the left maxillary sinus. A cyst may arise on any surface of the maxillary sinus. The cystillustrated in Fig. 51-14, B, arises from the anterior sinus surface. Cysts may be multiple orbilateral as demonstrated in Fig. 51-14, C.

The evaluation of the adjacent bony architecture is critical when attempting to assess theetiology of soft tissue pathologic findings. When dealing with a chronic inflammatory process,the adjacent bony architecture remains intact. If such an inflammatory process is chronic, anadjacent osteitis may ensure that can produce marked bony thickening (Fig. 51-15). In thepresence of a neoplastic process, adjacent bony erosion is usually present. The bony erosion ofneoplasia is usually aggressive compared to the erosion secondary to an expansile mass (usually

10

caused by mucoceles).

Mucosal thickening

Allergic or inflammatory insult may result in thickening of the mucous membrane aroundthe margin of a sinus wall to produce on radiographs a halo of increased density surrounding acentral air collection. Variable thickness characterizes mucosal changes, as demonstrated in Fig.51-16. The left maxillary sinus in Fig. 51-16, A, has thickening of the mucous membrane alongthe lateral and orbital surfaces. Diffuse bilateral mucosal thickening is present in both maxillarysinuses of the patient in Fig. 51-16, B. Only a very small central part of the sinus contains air.Tomography may indicate a lobular character to mucous membrane. The lateral tomogram in Fig.51-16, C, shows lobular mucous membrane along the anterior, alveolar, and posterior maxillarysurfaces. Similar lobular mucosal change in the sphenoid sinus is present on the coronaltomogram in Fig. 51-16, D.

Thickening of the mucous membrane is not often evident in the frontal sinus on plainfilms, probably because of the small volume of the frontal sinus. Mucosal thickening usuallyresults in complete opacity of the ethmoid sinus. The extent of mucosal thickening is bestdemonstrated by the CT examination (Fig. 51-17). The ability to show the adjacent bonyarchitecture and its relationship to the normal and pathologic mucosa continues to make CT theoptimal modality.

Borders

After evaluating sinus content, the examiner should turn to evaluation of the sinus wall.Several forms of change may be defined and represent a deviation from the normal thin laminadura representation of the normal osseous border.

Osteitis

Bone response may occur in relationship to an acute, high-grade infection that extendsthrough the haversian system of the sinus wall and produces lytic osteitis, in which bonesurrounding the infection seems to dissolve.

A well-defined region of bone absence resulting from lytic osteitis may be evident onradiographs, or infection of adjacent soft tissue surrounding the sinus wall may occur withoutspecific radiographic evidence of bone absence. In the latter case the area of bone lysis may betoo small to separate from uninvolved bone overlying it.

In response to low-grade, repeated infections, thickening of bone in the sinus wall mayresult in a buttressing increase of bone known as osteoblastic osteitis. Presumably osteoblasticosteitis is an effort of the sinus wall to isolate an intrasinus infection. Usually osteoblastic osteitisis quite evident on plain films as an area of bone wall thickening and increased density.Osteoblastic osteitis may occur in the wall of any sinus space. Lytic osteitis may not be evident

11

on special examination with tomography or high-resolution CT.

In Fig. 51-18, A, the right frontal sinus is opaque. The peripheral lamina dura in thelateral sinus border is missing, and the inferomedial (orbital) border is also absent. Lytic osteitisis present in these areas as well as in the area of the upper right ethmoidal lamina papyracea.Considerable soft tissue swelling of the periorbital soft tissue and lids reflects orbital cellulitison the right accompanying the lytic osteitis. The right maxillary sinus is also opaque, but noosteitis is present in this sinus wall.

The patient depicted by the CT scan in Fig. 51-18, B, had significant swelling of the rightcheek. The right maxillary sinus is opaque, and an area of lytic osteitis is present along theanterior and zygomatic surfaces of the sinus wall.

Osteoblastic thickening of the right frontal sinus accompanies sinus opacity in Fig. 15-19,A. Similar osteoblastic thickening is present along the right lateral maxillary wall. The CTexamination in Fig. 51-19, B, shows osteoblastic osteitis of the left maxillary sinus wallaccompanying thickening of the mucous membrane.

The coronal and lateral frontal sinus tomograms in Fig. 51-19, C and D, show right sinusopacity accompanied by distinct osteoblastic osteitis. No evidence of lytic change is present. Atoperation, hyperplastic mucous membrane filled sinus space, and microscopic examinationrevealed chronic inflammatory cells throughout the mucosa.

From the foregoing cases one can appreciate that plain films and tomograms may wellillustrate osteoblastic osteitis, but the sinus is opaque because of the overlying bone thickening.One cannot visualize the mucous membrane on plain radiographic studies when osteoblasticchange is present. However, the mucous membrane is well visualized in CT examination (Fig.51-19, B). Osteoblastic osteitis may also occur as a postoperative sinus change as described ina later section on surgical changes.

Bone proliferation

Proliferative bone thickening and overgrowth as found in fibrous dysplasia should not beconfused with osteoblastic osteitis. Not only does the sinus wall become markedly thickened, butalso the overall size of the sinus wall increases with proliferation. Commonly the alveolar portionof the maxilla is also enlarged and increased in density with the proliferative change of fibrousdysplasia.

The patient whose Caldwell's view is shown in Fig. 51-20, A, had a prominent right cheekand malar eminence. Also the right maxillary buccal surface was enlarged, and the teeth in thisarea did not articulate well with the mandibular dental structures. Significant thickening of theorbital and lateral maxillary walls existed, and the sinus air space was diminutive. A coronaltomogram of this patient (Fig. 51-20, B) showed the overall increase in size of the right maxillaas well as encroachment of the sinus. The appearance of bone in fibrous dysplasia has been

12

described as "ground glass", which certainly accurately describes the bone in this case.

Fibrous dysplasia producing blastic thickening in one portion of a sinus may be associatedwith expansile change in another part of the sinus. The expanded sinus wall is seen in associationwith an enlarging ossifying fibroma contained within the sinus cavity. This expansile change ofthe sinus wall should not be confused with the expansile changes produced by a mucocele.

The bone proliferation of fibrous dysplasia should not be confused with an osteoma, inwhich the margin of the dense bone mass is usually much more sharply defined.

Osteoma

Local areas of bone proliferation can occur anywhere in the skeleton. Histologically thesemasses are composed of an abnormal amount of dense cortical bone, although the cellularcomposition of the bone is normal. The frontal sinus is the most commonly involved paranasalsinus. The large left frontal osteoma in Fig. 51-21, A, obliterates and arises from the left sinusmargin. Expansion of the mass has produced slight depression in position of the subadjacentorbital roof. The lobular rounded border of the lesion helps differentiate an osteoma from fibrousdysplasia.

Osteomas are uncommon in the maxillary sinus. The osteoma illustrated in Fig. 51-21,B, is accompanied by osteoblastic change in the lateral sinus wall, which at least suggests anassociation with infection. Multiple sinus osteomas should suggest the possibility of Gardner'ssyndrome. Fig. 51-21, C, shows right frontal and ethmoidal sinus osteoma. The ethmoid sinusosteoma is more clearly defined on the coronal tomogram in Fig. 51-21, D. This patient also hadthe precancerous adenomatous polyps of the colon associated with Gardner's syndrome.

Expansion of sinus wall

An aerocele produces progressive enlargement of the involved sinus, which contains air.The sinus walls become very thin in association with such enlargement. An aerocele results fromone-way ingress of air into the sinus as a result of a "trap-door" change in the sinus wall and isusually the result of trauma (Zizmor and Noyek, 1978). This condition is very rare.

A much more common cause of expansile change is a mucocele. This cystic abnormalityis produced as a complication of inflammatory disease or by any other obstructing lesion suchas an osteoma. Posttraumatic mucoceles may occur if injury obstructs the sinus ostium. Typically,the mucoceles gradually enlarge, producing thinning of the sinus wall and obliteration of thenormal sinus shape. Usually the sinus is opaque because of the contents of the cystic lesion.

In Fig. 51-22, A, the right frontal sinus is opaque in comparison to the small left cell.Secondary septae around the sinus margin are absent and the central septum is displaced to theleft. The right medial orbital border is depressed and is so thin adjacent to the anterior ethmoidalcells that it seems to disappear. Fig. 51-22, B, is a lateral view of A. Expansile thinning of the

13

right frontal anterior and posterior walls is more evident. Least well visualized is the expansilechange of the involved anterior ethmoid sinus.

Ethmoid and sphenoid sinus involvement by a mucocele may require special examination.Tomographic study was necessary to demonstrate the left ethmoidal mucocele in Fig. 51-22, C.Plain radiographs of this patient demonstrated only ethmoidal opacity. The tomogram showedexpansile change of the left anterior ethmoidal cells and lamina papyracea, opacity of thesupraorbital cell, and thinning of the adjacent orbital roof.

Plain-film examination of the patient illustrated by Fig. 51-22, D, demonstrated onlysphenoid sinus opacity. The CT examination revealed expansile displacement of the sphenoidseptum along with sinus opacity.

Expansile change has also been described in association with primary fibrous tumors andneurofibromas that affect sinus walls (Bergeron et al, 1984). Expansile change may also occuraround the margins of a malignant sinus tumor, but the more specific change in sinus malignanttumor is erosion of bone.

Erosion of sinus wall

As tumor extends through the haversian system of the sinus bony wall, erosion results.The destructive change that tumor extension produces results in local or widespread absence ofthe sinus lamina dura. This change may be similar in appearance to that produced by osteoblasticosteitis in infections.

Waters' view was obtained to evaluate nasal obstruction in the patient depicted by Fig.51-23, A. In this study a small soft tissue mass is present along the lower right lateral maxillarywall that has an intact lamina dura. Sinus pain prompted a repeat examination of this patient 3years later. The second Waters' view (Fig. 51-23, B) revealed enlargement of the soft tissue masswith erosion of the overlying sinus wall. At surgery an adenoid cystic carcinoma was removed.

The more extensive undifferentiated right maxillary neoplasm demonstrated by thetomogram in Fig. 51-23, C, produced erosion of a large area of the nasal surface andpterygomaxillary surfaces. The ethmoidal, orbital, and anterior walls of the maxillary sinus arealso vulnerable to neoplastic erosion.

The CT scan in Fig. 51-23, D, showed involvement of the pterygoid muscles by amaxillary neoplasm. On the right posterolateral and nasal sinus wall, erosion was present. Thetumor invaded the right lateral pterygoid muscle and produced this mass.

CT best demonstrates the degrees of extension beyond the limits of a sinus primarilyinvolved by a malignant tumor. Although plain-film examination, as shown in Fig. 51-24, A,effectively demonstrates the bony erosion along the orbital border of the ethmoid sinus affectedby undifferentiated carcinoma, the CT scan in Fig. 51-24, B, better defines the extension of tumor

14

into the trochlea and superior rectus muscle. The tumor present in the right ethmoid sinus in Fig.51-24, C, has eroded through the posterior cell wall into the sphenoid sinus. Fig. 51-24, D, is acomposite CT study of a left maxillary sinus neoplasm that has eroded through the floor of theorbit and into the right ethmoid sinus. Erosion through the lamina papyracea produces lateraldisplacement of the eye and medial rectus muscle.

Clinical Applications

Infection

Nasal endoscopic and CT findings confirm that the ostiomeatal complex is the mostfrequent site of inflammatory sinus disease and that disease may persist in this area after thesecondary sinus disease has resolved. Nasal endoscopic examination consistently revealssignificant but subtle evidence of disease that is not visible on anterior rhinoscopy, to the extentthat the value of this diagnostic modality can no longer be seriously questioned. Further, the useof CT has allowed the incidence of concomitant anatomic variations and the extent of disease tobe accurately recorded. In a study of 230 patients with chronic sinus complaints, we found that78% had disease within the anterior ethmoid. The next most frequent site was the maxillary sinus(66%). The posterior ethmoid cells were involved in only 31%, a strong argument against thenecessity for a routine ethmoid dissection that begins posteriorly. Similarly, the frontal sinus wasonly involved in 34% of patients and the sphenoid in 16%. At the time of the study, 16% of thepatients had no evidence of inflammation. However, some of the patients had significantanatomic abnormalities that could possibly have been responsible for their prior recurrent disease.Therefore 93% of the patients with radiologic evidence of inflammation at the time of the studyhad disease within the anterior ethmoid. Evaluation of the same CT studies also revealed a widevariety of anatomic variations. The most common variation was concha bullosa, which occurredin 36%. The second most frequent finding was a septal deformity. Other variations occurred withdiffering frequencies.

Radiology and endoscopic surgery

The application of endoscopic and radiologic techniques for diagnosis and planning ofendoscopic surgical therapy introduced a radical new approach to sinus surgery. To apply thesetechniques successfully, however, a clear understanding of the pathophysiologic changesresponsible for the development of paranasal sinus inflammation is necessary.

Pathophysiology and mucociliary clearance

The middle turbinate area has been shown, by experiments using models, to be theprimary route involved in inspiratory airflow through the nose. Although the highest velocity ofairflow occurs through the area of the nasal valve, high velocity continues through the area ofthe anterior middle turbinate. This area then is exposed to a high velocity of unfiltered, poorlyhumidified air and bears the brunt of atmospheric pollutants, particulate matter, and bacteria inthe environment. The knowledge that adenocarcinoma in woodworkers originates in this area and

15

the results of experiments using tagged aerosol both support the concept of this being a key areafor the deposit of particulate matter from inspired air.

The key to the majority of inflammatory sinus disease is the middle meatus and anteriorethmoid (ostiomeatal complex). The ostiomeatal complex is composed of narrow channels andis subject to wide normal anatomic variation. It contains the ostia not only for the drainage ofthe anterior ethmoid sinuses but also for the frontal and maxillary sinuses. These sinuses thereforedepend on the integrity of the ostiomeatal complex for ventilation and mucociliary clearance.Therefore local inflammation or anatomic obstruction within the ostiomeatal complex may resultin secondary disease within the maxillary and frontal sinuses.

The ability of mucosal disease in the major sinuses to resolve, when normal mucociliarydrainage and ventilation are reestablished, is currently undergoing reevaluation. Clinicalexperience demonstrates that in many cases what had previously been considered irreversibledisease may recover when obstruction and inflammation are removed from the ostiomeatalcomplex.

CT examination, usually performed after medical treatment of sinusitis, can reveal theextent of mucosal disease deep in the ostiomeatal complex. Resolution of secondary inflammatorychanges affords a more accurate display of the regional anatomy and therefore provides a better"guide" for the surgical procedure.

The characteristics of chronic inflammatory disease on CT are mucoperiosteal thickening,"soft tissue mass", and osteitis of the ethmoid bony architecture. It is unusual to observe bonyerosion, because such erosion is most often associated with more invasive processes such as amucocele or a neoplastic process. In our experience an exception to this guideline is the presenceof erosion of the uncinate process, usually the result of chronic inflammation surrounding the"free edge" of this structure. This process, however, could also be explained as a mucocele ofthe maxillary sinus. The posterior ethmoid and the sphenoid sinuses are hidden from theendoscopist's direct view and access by the basal lamella. It is fortunate that an infection ispresent in this location in only 20% to 30% of the population with chronic inflammatory disease.When these air-containing spaces are blocked by inflamed mucosa, mucosal removal and drainagebecome possible under endoscopic guidance. In these cases, sagittally oriented reconstructionsfacilitate endoscopy and therapeutic instrumentation by affording a view of the plane assumedby the penetrating instruments. Distances and angulations using the anterior nasal spine as a pointof reference can be used to guide the endoscopist in the posterior advancement of the endoscope.

The information gained from CT evaluation is critical for presurgical planning. Asequential evaluation of the sinuses, ostia, and interconnecting air passages, starting anteriorlyand progressing posteriorly, should be made.

A systematic evaluation should include the frontal recess and adjacent agger nasi cell, theorientation of the middle meatus, and the infundibulum. The paradoxic shape of the middleturbinate should be addressed in relation to the middle meatus and infundibulum. Similarly, when

16

Haller cells are present, their possible infringement on the infundibulum should be considered.Because during endoscopic sinus surgery an uncinectomy is usually performed, the position ofthe "free edge" of the uncinate process in relation to the orbit needs to be clearly perceived bythe surgeon. In most instances the ethmoid bulla may be positioned between the uncinate processand orbit, but commonly this structure may directly adhere to the lamina papyracea. In the lattercase, care must be taken not to incise the medial orbital wall during this procedure.

In the majority of patients, the ethmoid bulla is separated from the sinus lateralis by anintact posterior wall. When this is the case, there may be inflammatory disease in the sinuslateralis in the presence of a normal ethmoid bulla. This information, only available to thesurgeon through CT, is critical if he or she intends to remove all the inflammatory pathologicprocess in the anterior ethmoid sinuses.

The presence of inflammatory disease in the posterior ethmoid and sphenoid sinusesnecessitates a display and understanding of the relationship between the width and height ofaeration of the posterior ethmoid sinus with respect to the sphenoid sinus. This relationshipstrongly influences the surgeon's choice of a "working plane" related to the position of the nasalseptum. Similarly, the extension of the carotid artery into the sphenoid sinus and its relationshipto the septations within this sinus must be noted to prevent a surgical complication. The decisionregarding surgical intervention should never be based on CT alone, because CT may demonstrateasymptomatic mucosal disease. Surgery is reserved for patients who fail medical management,and a decision to operate should be based on a combination of history, CT, and endoscopicfindings.

Future Directions

Improved imaging of the sinuses - especially the anterior ethmoid sinuses - is crucial foroptimal diagnosis and treatment of diseases of the nasal cavity and paranasal sinuses. Theincreased use of endoscopy and CT confirms the importance of ostiomeatal disease and anatomicdeformities of the middle meatus in the pathogenesis of sinusitis. Nasal endoscopy shouldsupplement clinical evaluation in all patients with chronic or recurrent acute sinusitis. CT shouldbe used when endoscopy fails to explain symptoms of sinusitis. CT shows the extent of mucosaldisease deep in the ostiomeatal complex and optimally displays the regional anatomy.

There is a need, however, to improve the three-dimensional perception of the complex andvarying anatomy of the nasal cavity and paranasal sinuses. Three-dimensional reconstruction ofdigitized CT data has been attempted but continues to be problematic given the very small airpassages, the fine mucosal lining, and very fine bony architecture (Fig. 51-25). The availablesegmentation algorithms need considerable improvement to become adaptable for this clinicalapplication. Nevertheless, the multiplanar two-dimensional reconstruction of the data, with asimultaneous display of axial, coronal, and sagittal images on which a particular anatomiclandmark may be identified in real time, is available and may be very useful for surgical planningand even for guiding the surgical procedure.

17

The equipment needed to accomplish this task consists of a computer capable of two-dimensional and three-dimensional reconstruction of CT and MRI data (ISG Technologies,Toronto), a high-resolution display screen, and a robotic mechanical sensor (FARO Medical,Florida) connected to the computer. This system was tested on a plastic replica of a skull andseveral cadaver head specimens. Five markers taped to the surface of the test subject were usedto register the subject position with the three-dimensional reconstruction on the computer'sdisplay screen.

Subsequently, an interactive correlation between anatomy on the test subject and its three-dimensional display was accurate within 0.7 to 1.8 mm. This development promises to improvethe endoscopist's perception of the anatomy in the operating field and the safety of this operativeprocedure. To date this development has been tested in two endonasal surgical procedures(ethmoid encephalocele and ethmoid tumor), and its accuracy proved to be similar to the accuracydetermined in in vitro testing.

Summary

Nasal endoscopy and the clinical evaluation form the basis for diagnosis of chronic andrecurrent sinusitis. Plain films offer no additional information in this setting. CT should be usedwhen endoscopy fails to explain symptoms of sinusitis. It should be performed after adequatemedical therapy, and, if possible, after administration of a local vasoconstrictor (to diminish theeffect of the nasal cycle). The milliamperage may be significantly lowered (as suggested byBabbel et al, 1991) to decrease the radiation exposure. A CT examination (preferably in thecoronal plane) should be available to the endoscopic surgeon for all surgical procedures. MRIfacilitates the diagnosis of nasal and paranasal sinus neoplasms and fungal disease. Computers,when coupled with a robotic mechanical sensor, improve anatomic perception and surgicalaccuracy.